Multidisciplinary design optimization of a non-planar flying wing UCAV concept

This study presents the multidisciplinary design optimization of a novel non-planar unmanned combat aerial vehicle geometry, aiming to maximize aerodynamic efficiency at a cruise condition of Mach 0.8 and an altitude of 11,000 m. The optimization adheres to strict constraints on aerodynamic performance (lift and drag), minimum airfoil thickness, and average radar cross-section. The latter is defined as the arithmetic mean of linear-scale values over all azimuthal angles at a 10 GHz frequency (X-band). The integrated multidisciplinary framework combines high-fidelity computational fluid dynamics and radar cross-section prediction solvers. Parametric non-planar lifting system geometry was generated using the Engineering Sketch Pad, while aerodynamic analysis and radar cross-section calculations were performed using the open-source SU2 suite and the POFACETS code, respectively. An in-house constrained efficient global optimization algorithm, based on Kriging surrogate modeling enhanced by the Partial Least Squares technique, was selected for its ability to efficiently navigate the high-dimensional, constrained design space with expensive function evaluations. This iterative process, built upon an initial dataset of 500 samples and refined through 100 additional high-fidelity evaluations, achieved a remarkable 44.55% improvement in aerodynamic efficiency compared to the baseline planar vehicle. This significant gain was realized by strategically managing the inherent aerodynamics-stealth trade-off, accepting a controlled increase in radar cross-section from the baseline’s 0.0005 m² to 0.0082 m² while satisfying all predefined constraints. Overall, this work demonstrates a comprehensive methodology for designing and optimizing non-planar unmanned combat aerial vehicles, effectively leveraging integrated open-source tools to deliver a high-performance and constrained multidisciplinary optimal solution.